When compiling content for the first reboot issue of Compute!’s Gazette, I knew one of the things that I wanted to feature was newly developed video games for our beloved classic hardware. I also wanted to feature nostalgic advertisements similar to how big box games used to be advertised in magazines: Full pages of glorious artwork depicting our heroes and villains in ways not able to be represented graphically by the technology of the time. To me, these ads always sparked my imagination and helped me get more invested in the game. I never felt let-down or had any negative feelings because the game itself didn’t represent the advertisement, because I knew the hardware limitations and never expected it to.

I consider myself a decent artist who at least has an eye for design, and given enough time and inspiration could manually draw up decent artwork for something like a game cover or advertisement. However, I’m a single individual person and have taken on a huge undertaking trying to revive Compute!’s Gazette magazine mostly on my own for now. And while I do have plenty of inspiration, I don’t have a lot of time. I am also a general technologist, and generally speaking embrace new technology – so when it comes to using tools available to me such as Generative AI, I tend to embrace their use where it makes sense. While I wouldn’t even consider blindly auto generating content using generative AI for example, I would maybe use it to proofread my work or suggest different perspectives. The same goes for artwork, I feel my strong points are layout and design and I periodically use Generative AI to help me realize ideas graphically, similar to the way I might produce a physical representation of an idea with a 3D printer.

Given the amount of work it takes and the resources I have available to me, there is simply no way I could expect to successfully achieve something like publishing a monthly magazine without using any and all resources available to me.

So after spending several hours of my weekend on both Saturday and Sunday designing what I thought to be an awesome looking advertisement for a DOS Freeware game that was representative of what I would expect to see had this magazine been published in 1991, I shared this advertisement with the developer of the game fully expecting a “WOW!” reaction.

The “WOW!” reaction was unfortunately from me. “WOW!” was I disappointed in the response I got back from the developer several hours later. Needless to say, the developer very directly and coldly informed me that, “Sorry, I don’t support generative AI for uses like these. In my opinion devalues the product.”

While it’s true that I didn’t develop the entire advertisement from scratch, the inverse is also true – it wasn’t completely the result of Generative AI either. I had actually played the game I was advertising, I had a picture in my mind of what I wanted and spent a couple of hours adjusting prompts and different Generative AI models until I was able to get a background image for the advertisement that I was happy with, and I spent a few hours on manual image manipulation, layout, and word art. The finished advertisement checked all the boxes I had for my vision and it was something I would have fully expected to see in a 1991 issue of Compute!’s Gazette. Had I produced it in 1991 and given it to a client, I have no doubt in my mind that they would be blown away by it. I was proud of my creation, and now I was made to feel as if what I had created was worthless and meaningless and otherwise garbage, despite putting significant creative energy into it.

I immediately went through the 5 stages of grief in succession and in very short order: Denial, Anger, Bargaining, Depression, and Acceptance. I mean, the advertisement I produced was a result of MY vision and MY work but was he right? Was it meaningless simply because I used advanced tools to create it? By the time I got to acceptance, it was clear to me that the developer had very strong opinions about the use of Generative AI, and I wanted to understand the reasoning behind those feelings.

As a result, I’ve decided to not publish the artwork I had produced out of respect for his feelings about the matter.

Nevertheless, I was intrigued by his perspective and decided to dig a bit more with this particular developer on the subject and was pretty quickly shut down. Further attempts to dive deeper about acceptable uses of Generative AI in game development were immediately shut down as “off-topic”.

Not deterred, I decided to do a general poll of game developers. I belong to an Indie Game Developers group on Facebook and asked where developers fell in a spectrum of using Generative AI in game development – from “Will never use AI under any circumstances for game dev” to “Full use – will use whatever is available”

Some custom responses were also added by respondents – ranging from good to silly. “Not against using AI but haven’t found a viable use case” got 2% responses, and “Only use it when i’m poor when I hit big will hire artists” also got 2%. Some devs were even upset that AI was even brought up while others argued about how they are “developer gods” and could never be replaced by AI. One commenter, an artist, said “I’m an artist and being able to “somewhat” make projects without expensive plugins or help from [somebody] else is a game changer, all that hate most probably comes from those who are being replaced”

“guys, you’re acting like it was the end of the world for typewriters when printers were introduced.” one commenter mused.

Overall, I was pretty surprised by the results to be honest, especially from 15% of game developers stating they would never use AI under any circumstances. Game developers literally helped pioneer AI.

Another developer commented, “That reminds me of a carpenter who won’t use a nail gun. Then, again, I am not and never will qualify as a Luddite given that I have spent the entirety of a long life building the cutting edge of technology. Why would I not use tech I once tried to build? INANE!“

Inane indeed.

Remember the video game Elite? Written and developed by David Braben and Ian Bell, it was originally published by Acornsoft for the BBC Micro and Acorn Electron computers in September, 1984 and ported to a multitude of classic systems. My first inclination of what “AI” could do for game development came when I first played it on my C64 and wondered what sort of black magic was being used to put all that game inside of 64k of memory (I didn’t refer to it as AI back then, simply black magic or sorcery :)).

Elite is one of a few early games that often come up as examples of procedural generation. Elite, for those of you who haven’t played it, is a highly acclaimed space trading game simulation. You fly your ship from star to star, buying and selling various commodities. Those star systems were procedurally generated from a fixed seed, which generated the description of the star system, including its economy, government, and population. There were 8 galaxies of stars in the original Elite, each with 256 generated stars. This procedural generation is a very early, albeit more rudimentary, form of what we now understand as Generative AI.

If we were able to travel back in time and take this amazing tool called Generative AI with us to the headquarters of any major game studio – do you really think game devs at that time would throw us out of the building and tell us to bury this demonic technology? Not a chance.

So, why would we even consider that today? Moral reasons? I think I understand most of the potential moral arguments and I’m sure that they are mostly legitimate from a certain perspective, but also think that a lot of it comes from fear, uncertainty, and doubt. There are also equally compelling counter arguments to those moral reasons. Generally speaking, it reminds me of the stories of villagers with pitchforks coming to rid the village of demonic witches. I don’t believe that I’ve every witnessed anything that was done out of fear being worthwhile.

“Any sufficiently advanced technology is indistinguishable from magic.” -Arthur C. Clarke.

So lets explore a couple of the potential moral reasons we should consider not using Generative AI for game development:

“AI Will Replace Coders/Artists/ETC”

The beauty of a tool like Generative AI is it’s ability to be prompted for a result, produce a result, and refine that result based upon the creators preferences. Coders and Artists don’t just create things for no reason, and you must be able to at least think like someone who would traditionally create whatever it is you’re trying to create. If you don’t know anything about coding, do you really expect that you would be able to prompt a Generative AI to write code for “the coolest virtual reality game ever made” or “create an image of a painting more beautiful than the Mona Lisa” and it would produce something worthwhile and usable? Of course not.

Any developer who’s had a game pitched to them by one of their friends trying to describe “the coolest virtual reality game ever made” would immediately understand that while his/her friend might think that they have the clearest picture of what that game is in their minds eye, being able to clearly convey what that means is another story. Generally, these pitches fall apart pretty quickly once they come under scrutiny from a developer. Why would we expect that suddenly coders and artists would suddenly be out of jobs because a powerful tool exists that would allow Joe or Jane Blow the opportunity to AI generate unrefined raw code or art based on uneducated prompts? Regardless of how far the technology progresses, the need for clear requirements and expectations of results, and the knowledge of how to process and refine those results, will never go away.

The heart of the true value of a coder or artist is imagination, which will never be replaced by a machine. While talent can be learned and procedurally generated such as writing clean code or tactful manipulation of a drawing instrument, the application of that talent is completely up to the imagination of the human creator. Perhaps imagination can be trained and somewhat simulated but it will never be completely reproduced. Quality output will always be dependent upon quality input. Conversely, garbage in = garbage out.

“AI Steals Copyrighted Work”

Generative AI models are trained on vast datasets, often scraped from the Internet. This data frequently includes copyrighted material like images, text, music, and code.

The concern from many is that using this copyrighted material without explicit permission or licensing constitutes infringement on the original creators of this material.

The problem I see with this argument is that human beings are also trained on vast datasets, often scraped from the Internet, which also contains copyrighted material like images, text, music, and code. This is generally how we learn, and anyone who lives and creates today isn’t producing anything completely new without knowledge or influence from another creator. Artists often imitate the style and content of their favorite artists. The same goes for musicians and coders. Unless you grew up isolated in a laboratory with no access to the outside world, anything you produce today is likely largely influenced by someone else’s work in some form or fashion. Anyone whose listened to Greta Van Fleet should immediately recognize the band’s influence. That’s not a knock to the band at all, I thoroughly enjoy their music – and it’s completely original music. It’s just a good illustration of human creative influence.

Does working under the influence of others make your work immoral?

In conclusion, while the prospect of AI Overlords taking over and relegating all of humanity to the unemployment lines is a scary one, it’s fictional. Can AI make things more efficient and effectively replace some roles? Sure it can – that’s the point. But that’s the role of technology in general and those relegated to the bread lines are those who are unwilling to adapt to the changing of the times. Technology itself is still subject to the laws of supply and demand. There should never be a demand for unimaginative works that are bland and devoid of soul, therefore there will likely never be a soulless AI independently supplying a nonexistent demand.

Remember the past but embrace the future. Everything old is new again.

Did you enjoy this article? Click here to stay informed, support retro, and subscribe to Computes Gazette today!